Perceptual Manipulations for Hiding Image Transformations in Virtual Reality

Users of a virtual reality make frequent gaze shifts and head movements to explore their surrounding environment. Saccades are rapid, ballistic, conjugate eye movements that reposition our gaze, and in doing so create large-field motion on our retina. Due to the high speed motion on the retina during saccades, the brain suppresses the visual signals from the eye, a perceptual phenomenon known as the saccadic suppression. These moments of visual blindness can help hide the display graphical updates in a virtual reality.

In this dissertation, I investigated how the visibility of various image transformations differed, during combinations of saccade and head rotation conditions. Additionally, I studied how hand and gaze interaction, affected image change discrimination in an inattentional blindness task. I conducted four psychophysical experiments in desktop or head-mounted VR. In the eye tracking studies, users viewed 3D scenes, and were triggered to make a vertical or horizontal saccade. During the saccade an instantaneous translation or rotation was applied to the virtual camera used to render the scene. Participants were required to indicate the direction of these transitions after each trial. The results showed that type and size of the image transformation affected change detectability. During horizontal or vertical saccades, rotations along the roll axis were the most detectable, while horizontal and vertical translations were least noticed. In a second similar study, I added a constant camera motion to simulate a head rotation, and in a third study, I compared active head rotation with a simulated rotation or a static head. I found less sensitivity to transsaccadic horizontal compared to vertical camera shifts during simulated or real head pan. Conversely, during simulated or real head tilt observers were less sensitive to transsaccadic vertical than horizontal camera shifts. In addition, in my multi-interactive inattentional blindness experiment, I compared sensitivity to sudden image transformations when a participant used their hand and gaze to move and watch an object, to when they only watched it move. The results confirmed that when involved in a primary task that requires focus and attention with two interaction modalities (gaze and hand), a visual stimuli can better be hidden than when only one sense (vision) is involved.

Understanding the effect of continuous head movement and attention on the visibility of a sudden transsaccadic change can help optimize the visual performance of gaze-contingent displays and improve user experience. Perceptually suppressed rotations or translations can be used to introduce imperceptible changes in virtual camera pose in applications such as networked gaming, collaborative virtual reality and redirected walking. This dissertation suggests that such transformations can be more effective and more substantial during active or passive head motion. Moreover, inattentional blindness during an attention-demanding task provides additional opportunities for imperceptible updates to a visual display.